HPCwire

Leading HPC
Solution Providers




















HPCwire >> Blogs

Blog: From the Editor

From the Editor | Main Blog Index

SiCortex Gets Personal


Has SiCortex found the right formula for the personal supercomputer? Introduced in November 2007, the company's Catapult SC072 is a deskside mini-cluster that can be plugged into a standard wall outlet. Positioned as the entry-level system in the SiCortex family of MIPS processor-based supers, the Catapult has a modest 72 processors, 48 GB of memory, 2 gigabit Ethernet ports, and up to 3 PCI Express cards, each operating at a peak speed of 2 GB/sec. The enclosure has enough room for up to six standard disk drives. Oh, and it looks cool.

The SC072 system tops out at about 70 peak gigaflops -- slightly less than you could achieve with a two socket machine based on the latest quad-core Intel Xeon processors. A Xeon-based system has the additional advantage of only requiring 8-way parallelism (two sockets of quads) from the application. The Catapult requires the software to achieve 72-way parallelism to keep all the computational engines busy. But the SiCortex offering is built with the cluster/MPI crowd in mind, and 72 processors is fairly modest by today's standards. Where the Catapult really shines is its low power consumption (less than 200 watts), its relatively large memory (48 GB), and the ability to include a RAID disk. Add to that the integrated software development environment, and a suggested price of under $15K, and you have a very interesting entry-level machine for HPC application developers.

"We want to explode the myth that SiCortex -- unlike some other companies -- is only selling multi-million dollar systems, and that the technology is for the rich and famous," explains SiCortex CEO John Mucci. "The ISVs are terrified of these kinds of computers. Even though this machine [Catapult] has a different instruction set, fundamentally it's a Linux-MPI-C/Fortran -- someday UPC -- computer. But you don't need to mortgage your soul or go the the facilities people to get one."

The company views the Catapult as synergistic with their larger offerings -- the SC648, SC1458, SC5832 -- and has developed a three-pronged strategy around the SC072:

  • First is the idea of 7/24 availability for HPC developers. Because of its personal nature, Catapult provides a perfect development platform for SiCortex customers with larger systems. As any HPC developer knows, getting allocations on a larger time-shared machine constrains iterative software development. Much of the software development process, HPC or otherwise, is still very intimate, so the availability of a personal computing platform offers a tremendous advantage in software productivity.

  • Next is getting the meme established that these systems are not "special" inasmuch as you can have a multi-thousand dollar system that works essentially the same as a multi-million dollar system. Having an architecturally consistent set of offerings, whose names conveniently reflect the machine's processor count, make the product line simple to understand.

  • Finally, there's the notion that a personal supercomputer can be a powerful educational tool for parallel programming -- at least in the MPI sense. At less than $15,000 a pop, there's an opportunity to get a lot of these systems into the hands of commercial developers and CS students. As is stands today, there's still only a small population of programmers who understand parallelism and part of that has to do with a simple lack of access to real HPC hardware. According to SiCortex, some of their customers are interested in proliferating these machines into the academic community.

Mucci says those strategic imperatives are more than justification for having built the Catapult. And while he admits there's no sales force in place dedicated exclusively to the SC072, it's quite possibile that the product will be a moneymaker on its own. At this point, only a few organizations are purchasing the small machines separately. According to Mucci, they're seeing a strong pipeline for their larger systems and they'll be shipping one or more Catapults with about 3/4 of those orders.

Early adopters are already running some codes on the machines. Rutgers got the first Catapult in January. They would like to have a whole lab full of them to help teach computer science and engineering. For the time being, students in the spring 2008 course on parallel computing will make use of the single system. The Rutgers system will also be used to develop algorithms to study mitochondrial and viral evolution, and to find markers for cancer treatment.

This week it was announced that Jack Dongarra at the Innovative Computing Lab, University of Tennessee, has acquired a Catapult. There it will be used by students to learn about and develop HPC software. Dongarra's group will also use the system to devise a new parallel processing communication performance benchmark and port and tune the OpenMPI communication library.

At Argonne National Lab, Rick Stevens, associate director of Computing and Life Sciences, also has a Catapult in his office. The system was purchased in conjunction with the 5.8 teraflop SC5832 the lab purchased in October 2007. No word on what Stevens is up to with his system, but since he still dabbles in programming, one can assume his personal SC072 is getting a workout.

SiCortex has managed to get the system into the price range that allows buyers to bypass institutional purchasing processes. Recently, a small ISV that develops HPC application software bought a Catapult so they'd have a platform to port their code to the MPI/cluster model. The price meant the CEO was able to just put one on his personal account and hand it over to his developers.

While the Catapult is primarily intended as a development platform, the size of the system makes sense for certain end-user applications. Boston University Medical Center is using an SC072 to run experiments with patient data from the Framingham Study, with the intent of correlating lifestyle, diet, and physical characteristics with heart disease. The ability to have a compact high-powered system that can be secured easily (because of patient privacy issues) was a good fit for this particular application. And although the Catapult was the right size for this particular work, the medical center might end up buying additional systems or even a larger machine.

In some ways, SiCortex is starting to look like the Apple Computer of the HPC world. Like Apple (up until 2006), SiCortex eschewed the x86 in favor of a more efficient processor to power its computers. Also like Apple, SiCortex developed a more highly integrated, customized system architecture than its more mainstream competitors. If SiCortex could find a way to spread the Catapult into the academic community, it would again be following in Apple's footsteps. While the Mac OS-based products only have a six percent US market share, that's still a nice slice of the pie. SiCortex would be more than happy to get that cut of the $10 billion-plus HPC server market. Now, if the company could just come up with the HPC equivalent of the iPod, they'd be set.

-----

As always, comments about HPCwire are welcomed and encouraged. Write to me, Michael Feldman, at editor@hpcwire.com.

Posted by Michael Feldman - April 11 @ 12:00AM

(Digg, Technorati, more)

Discussion

There are 0 discussion items posted.  

Sponsored Links

New Paper: Parallel Computing Without Parallel Programming
Learn how domain experts can run VHLL programs like MATLAB® on a variety of high-performance platforms without low-level reprogramming and how to work with the largest datasets and complex algorithms without sacrificing ease of use or reducing productivity.

Michael Feldman

Michael Feldman is the editor of HPCwire.

More Michael Feldman



Recent Comments

Feature Articles

Spider Up and Spinning Connections to All Computing Platforms at ORNL

Spider, the world's biggest Lustre-based, centerwide file system, has been fully tested to support Oak Ridge National Laboratory's new petascale Cray XT4/XT5 Jaguar supercomputer and is now offering early access to scientists.
Read More...

Wolfram Alpha: A Web-Based Application That Embraced Supercomputers

Wolfram Alpha, the Web-based computational engine introduced in May, is not a traditional supercomputing application, but relies on supercomputers to satisfy its unique requirements.
Read More...

TeraGrid '09: Student Participation Soars

There was a new energy at this year's TeraGrid '09 conference thanks to an outstanding turnout for the student program. Thanks to support from the National Science Foundation, more than 100 high school, undergraduate and graduate students were able to participate in the conference.
Read More...

Top Headlines

3D Seismic Data: Taking a Smarter Approach to Interpretation

Jul 09 | Engineer Live | The demand for computational tools to underpin the 3D seismic interpretation process has never been more apparent. Read more...

Engineering Unemployment Soared in 2Q to 8.6%

Jul 08 | EE Times | Unemployment for U.S. engineers has reached record levels, according to government figures. Read more...

Gartner Adjusts 2009 IT Spend Downward Again

Jul 08 | Network World | Global spending for 2009 projected to drop 6 percent, for a total of $3.2 trillion. Read more...

Concurrent and Parallel Are Not The Same

Jul 08 | Linux Magazine | Portability or efficiency? Neither is guaranteed when writing explicit parallel code. Read more...

800 TFLOP Real-Time Ray Tracing GPU Unveiled, Not for Gamers

Jul 07 | Ars Technica | Japanese company builds custom ASIC to accelerate real-time ray traced rendering for the auto industry. Read more...

Featured Whitepapers

Parallel Computing Without Parallel Programming

Jul 10 | | Engineers, scientists, and other domain experts depend on the productivity enabled by very high-level language (VHLL) tools like MATLAB® and Python. However, as datasets grow larger and programs get more sophisticated, ordinary desktop computers can no longer keep up. The paper explores how to run VHLL programs on high-performance platforms without low-level reprogramming. Work with large datasets and complex algorithms without sacrificing ease of use or reducing productivity.

Building High Performance Computing in a Green and Modular Solution Building Block

Apr 14 | | Many HPC IT departments are feeling the rising pressure to deliver more capacity computing and performance while trying to reduce the total cost of ownership. This white paper discusses how an environmentally-friendly and open-standards HPC building block based computing system using flexible interconnect options helps address capacity computing needs.

Multimedia

Webcast: Dell Expands HPC Access and Adoption with Intel Cluster Ready Program


Source: Addison Snell, GM/VP, Tabor Research; sponsored by Dell

Many organizations that could benefit from the use of HPC clusters find that it is complicated to get the systems up and running because of limited IT resources or the complexities of the clusters themselves. Learn how the Intel Cluster Ready program, for which Dell was an original partner, seeks to address this challenge for entry level and mid-range HPC users.

Video White Paper: Architecting a Better Network Storage Solution

BlueArc's Titan architecture represents an evolutionary step in file servers by creating a hardware-based file system that can scale bandwidth, IOPS, and overall data capacity well beyond conventional software-based devices. With its ability to virtualize a massive storage pool of up to four usable petabytes of tiered storage, Titan can scale with growing data requirements, offering a competitive advantage for businesses, researchers, or other enterprises seeking to better manage data growth while still ensuring optimal performance.

Webcast: HPC Development Solutions: Sun Studio & Sun HPC ClusterTools


Sun Studio Compilers and Tools and Sun HPC ClusterTools allow you to create high performance parallel applications for OpenSolaris, Solaris and Linux. Sun Studio Express 11/08 includes MPI performance analysis capabilities and full OpenMP 3.0 compiler support. Learn about all this and the latest in Sun HPC ClusterTools 8.1.

Blogs by Topics

Blogs by Author

HPC Blogroll



Featured Events

WORLDCOMP 2009
Data Mining Courses